27 December 1996 Input image spectral density estimation for real-time adaption of correlation filters for optical pattern recognition
Author Affiliations +
Proceedings Volume 2969, Second International Conference on Optical Information Processing; (1996) https://doi.org/10.1117/12.262628
Event: Second International Conference on Optical Information Processing, 1996, St. Petersburg, Russian Federation
Abstract
The problem of image noise estimation for improved noise robustness and discrimination capabilities of optical correlation filters is discussed. Colored noise is often used in the literature as an approximation to the true noise spectral density in the input image of a correlator. This conjecture is verified on different kinds of input images, i.e. their power spectral densities are fitted to a colored noise model. The quality of the resulting approximation is discussed. It is then shown that incorporating this noise estimation into optimal trade-off filters can significantly improve both the discrimination capabilities and the signal to noise ratio of the resulting adaptive correlation filter above that of the classical filters for which the noise parameters are not estimated. Although its performances are in general found to be markedly inferior to those of true nonlinear filtering techniques that are optimal for adaptive image correlation, the proposed adaptive method is attractive in terms of computation time. The optical implementation of the proposed method is also presented.
© (1996) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Anders Grunnet-Jepsen, Anders Grunnet-Jepsen, Sylvie G. Tonda, Sylvie G. Tonda, Vincent Laude, Vincent Laude, } "Input image spectral density estimation for real-time adaption of correlation filters for optical pattern recognition", Proc. SPIE 2969, Second International Conference on Optical Information Processing, (27 December 1996); doi: 10.1117/12.262628; https://doi.org/10.1117/12.262628
PROCEEDINGS
6 PAGES


SHARE
Back to Top